ResNet50_Swish_CIFAR10_torch.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "7628cb51-ec9d-4d38-adcb-f36678bde2ee",
   "metadata": {},
   "source": [
    "# ResNet-50: CIFAR-10\n",
    "\n",
    "- Trained with Swish activation function: $f(x) = \\frac{x}{1 + e^{-\\beta \\cdot x}}$, where $\\beta$ is a trainable hyper-parameter. Or, $f(x) = x \\cdot sigmoid(\\beta \\cdot x)$.\n",
    "- Linear learning-rate warmup, followed by step-decays."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "afa812dd-961b-4ce1-a422-cba4c98802b3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "env: CUDA_DEVICE_ORDER=PCI_BUS_ID\n",
      "env: CUDA_VISIBLE_DEVICES=0, 1\n"
     ]
    }
   ],
   "source": [
    "# Specify GPU to be used-\n",
    "%env CUDA_DEVICE_ORDER = PCI_BUS_ID\n",
    "'''\n",
    "Specify GPU(s) to be used here. Say you have 6 GPUs and you want to train Model A on 2 of them\n",
    "and Model B on 4 of them. You can do that as follows-\n",
    "CUDA_VISIBLE_DEVICES=0,1 python model_A.py\n",
    "CUDA_VISIBLE_DEVICES=2,3,4,5 python model_B.py\n",
    "'''\n",
    "# %env CUDA_VISIBLE_DEVICES = 0, 1, 2, 3\n",
    "%env CUDA_VISIBLE_DEVICES = 0, 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "812af13d-b8da-4d23-b681-30acc2dfe249",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "\n",
    "from resnet_bottleneck_block import resnet_block, ResNet50\n",
    "from cifar10_dataloader import get_cifar10_data\n",
    "from tqdm import tqdm\n",
    "from tqdm import trange\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os, pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d36f79d8-754c-4252-97b9-7cad852a5ba1",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9cdcd502-a70d-41a3-b7d9-2d4eacbffe1f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bb58e0e3-3ee5-4404-9de4-828b3ea81536",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch version: 1.13.1\n",
      "Number of GPU(s) available = 2\n",
      "Current GPU: 0\n",
      "Current GPU name: NVIDIA TITAN RTX\n",
      "Available device is cuda\n"
     ]
    }
   ],
   "source": [
    "print(f\"torch version: {torch.__version__}\")\n",
    "\n",
    "\n",
    "# Check if there are multiple devices (i.e., GPU cards)-\n",
    "print(f\"Number of GPU(s) available = {torch.cuda.device_count()}\")\n",
    "\n",
    "if torch.cuda.is_available():\n",
    "    print(f\"Current GPU: {torch.cuda.current_device()}\")\n",
    "    print(f\"Current GPU name: {torch.cuda.get_device_name(torch.cuda.current_device())}\")\n",
    "else:\n",
    "    print(\"PyTorch does not have access to GPU\")\n",
    "\n",
    "# Device configuration-\n",
    "device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
    "print(f'Available device is {device}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a1656e60-ae43-407e-8303-bbd4214e0778",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "d9477170-5b2c-447a-bb1d-c88fac7423de",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Specify hyper-parameters\n",
    "batch_size = 256\n",
    "num_classes = 10\n",
    "num_epochs = 85"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd9ff2a9-5599-4f9d-ab21-9eef6194b842",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "31d51cd0-bece-4fcd-89e2-6358b61e0e5a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Files already downloaded and verified\n",
      "Files already downloaded and verified\n"
     ]
    }
   ],
   "source": [
    "# Get CIFAR-10 dataset-\n",
    "train_loader, val_loader, train_dataset, val_dataset = get_cifar10_data(\n",
    "    path_to_files = '/home/majumdar/Downloads/.data/',\n",
    "    batch_size = 256\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "df813694-884e-455d-8bac-eb79aa34f240",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "f4204ac3-1bfa-4e97-9b0a-9e52a8c21584",
   "metadata": {},
   "source": [
    "#### ResNet-50 + $\\beta$ parameter (Swish) from initial experiment\n",
    "\n",
    "A previous experiment found that the optimal $\\beta = 5.354$ gave the highest validation accuracy. So, a new ResNet-50 is initialized with this $\\beta$ value and retrained from scratch to compare the benefits of initializing the same network with the previously obtained optimal value."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5eb22e60-1c05-49e1-9b2b-a64b08c540a6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instantiate ResNet-50 CNN-\n",
    "model = ResNet50(beta = 5.354).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b7b8a08f-641a-4e07-a253-2cec86fad5fc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "beta is initialized as:5.3540\n"
     ]
    }
   ],
   "source": [
    "# Sanity check-\n",
    "print(f\"beta is initialized as:{model.swish_fn.beta.detach().item():.4f}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0447665e-2a66-4239-a057-f32bf211d502",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "cdc6e70e-e454-4528-a382-446463ade37e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def count_trainable_params(model):\n",
    "    # Count number of layer-wise parameters and total parameters-\n",
    "    tot_params = 0\n",
    "    for param in model.parameters():\n",
    "        # print(f\"layer.shape = {param.shape} has {param.nelement()} parameters\")\n",
    "        tot_params += param.nelement()\n",
    "\n",
    "    return tot_params\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "972aeac1-f780-4d47-be9e-1ca1109eeb2a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ResNet-50 CNN has 23520843 params\n"
     ]
    }
   ],
   "source": [
    "print(f\"ResNet-50 CNN has {count_trainable_params(model)} params\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b1420a0b-9602-4e93-b0e1-ad11719f5904",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f8f58956-7945-430b-8210-78076924dabe",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define loss function and optimizer-\n",
    "loss = nn.CrossEntropyLoss()\n",
    "\n",
    "optimizer = torch.optim.SGD(\n",
    "    params = model.parameters(), lr = 0.0,\n",
    "    momentum = 0.9, weight_decay = 5e-4\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2b039d39-48fd-44a6-9ca4-470110577739",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "602ad582-04d4-4a9a-8c05-713e68710ff1",
   "metadata": {},
   "source": [
    "#### Train model with learning rate scheduler\n",
    "\n",
    "Training dataset = 50000, batch size = 256, number of training steps/iterations per epoch = 50000 / 256 = 195.3125 = 195\n",
    "\n",
    "After an initial linear learning rate warmup of 13 epochs or 2539 training steps:\n",
    "\n",
    "- For the next 7 epochs, or, until $20^{th}$ epoch, use lr = 0.1.\n",
    "\n",
    "- For the next 25 epochs, or, until $45^{th}$ epoch, use lr = 0.01.\n",
    "\n",
    "- For the 25 epochs, or, until $70^{th}$ epoch, use lr = 0.001.\n",
    "\n",
    "- For remaining epochs (10 epochs), use lr = 0.0001."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c42074f6-7dda-4ee3-9414-2c1d5291d6ca",
   "metadata": {},
   "outputs": [],
   "source": [
    "boundaries = [3906, 8790, 13672]\n",
    "values = [0.1, 0.01, 0.001, 0.0001]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f01a2460-1fc2-4868-b76c-d43cbbe732d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "def decay_function(\n",
    "    step, boundaries = [3906, 8790, 13672],\n",
    "    values = [0.1, 0.01, 0.001, 0.0001]\n",
    "):    \n",
    "    for idx, bound in enumerate(boundaries):\n",
    "        if step < bound:\n",
    "            return values[idx]\n",
    "\n",
    "    return values[-1]\n",
    "\n",
    "class schedule():\n",
    "\n",
    "    def __init__(self, initial_learning_rate = 0.1, warmup_steps = 1000, decay_func = None):\n",
    "        self.initial_learning_rate = initial_learning_rate\n",
    "        self.warmup_steps = warmup_steps\n",
    "        self.decay_func = decay_func\n",
    "        self.warmup_step_size = initial_learning_rate/warmup_steps\n",
    "        self.current_lr = 0\n",
    "\n",
    "    def get_lr(self, step):\n",
    "        if step == 0:\n",
    "            return self.current_lr\n",
    "        elif step <= self.warmup_steps:\n",
    "            self.current_lr+= self.warmup_step_size\n",
    "            return self.current_lr\n",
    "        elif step > self.warmup_steps:\n",
    "            if self.decay_func:\n",
    "                return self.decay_func(step)\n",
    "        else:\n",
    "            return self.current_lr\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "8f587445-b39b-47c6-aa60-63035dcd5975",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initial linear LR warmup: 195.3125 x 13 = 2549 steps in 13 epochs.\n",
    "custom_lr_scheduler = schedule(\n",
    "    initial_learning_rate = 0.1, warmup_steps = 2539,\n",
    "    decay_func = decay_function\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "172d0880-3eab-42e3-b86d-c5d3c40ebfb6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "2f31fce2-adf0-4be1-9603-3060c6641a9e",
   "metadata": {},
   "outputs": [],
   "source": [
    "step = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "5cb7817c-89f2-4c7c-af70-ac614d21be3c",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model_progress(model, train_loader, train_dataset):\n",
    "    '''\n",
    "    Function to perform one epoch of training by using 'train_loader'.\n",
    "    Returns loss and number of correct predictions for this epoch.\n",
    "    '''\n",
    "    running_loss = 0.0\n",
    "    running_corrects = 0.0\n",
    "    \n",
    "    model.train()\n",
    "    \n",
    "    with tqdm(train_loader, unit = 'batch') as tepoch:\n",
    "        for images, labels in tepoch:\n",
    "            tepoch.set_description(f\"Training: \")\n",
    "            \n",
    "            images = images.to(device)\n",
    "            labels = labels.to(device)\n",
    "            \n",
    "            # Get model predictions-\n",
    "            outputs = model(images)\n",
    "            \n",
    "            # Compute loss-\n",
    "            J = loss(outputs, labels)\n",
    "            \n",
    "            # Empty accumulated gradients-\n",
    "            optimizer.zero_grad()\n",
    "            \n",
    "            # Perform backprop-\n",
    "            J.backward()\n",
    "            \n",
    "            # Update parameters-\n",
    "            optimizer.step()\n",
    "            \n",
    "            global step\n",
    "            optimizer.param_groups[0]['lr'] = custom_lr_scheduler.get_lr(step)\n",
    "\n",
    "            step += 1\n",
    "            \n",
    "            # Compute model's performance statistics-\n",
    "            running_loss += J.item() * images.size(0)\n",
    "            _, predicted = torch.max(outputs, 1)\n",
    "            running_corrects += torch.sum(predicted == labels.data)\n",
    "            \n",
    "            tepoch.set_postfix(\n",
    "                loss = running_loss / len(train_dataset),\n",
    "                accuracy = (running_corrects.double().cpu().numpy() / len(train_dataset)) * 100\n",
    "            )\n",
    "            \n",
    "    \n",
    "    train_loss = running_loss / len(train_dataset)\n",
    "    train_acc = (running_corrects.double() / len(train_dataset)) * 100\n",
    "    \n",
    "\n",
    "    # return running_loss, running_corrects\n",
    "    return train_loss, train_acc.cpu().numpy()\n",
    "\n",
    "def test_model_progress(model, test_loader, test_dataset):\n",
    "    total = 0.0\n",
    "    correct = 0.0\n",
    "    running_loss_val = 0.0\n",
    "\n",
    "    with torch.no_grad():\n",
    "        with tqdm(test_loader, unit = 'batch') as tepoch:\n",
    "            for images, labels in tepoch:\n",
    "                tepoch.set_description(f\"Validation: \")\n",
    "                \n",
    "                images = images.to(device)\n",
    "                labels = labels.to(device)\n",
    "                \n",
    "                # Set model to evaluation mode-\n",
    "                model.eval()\n",
    "            \n",
    "                # Predict using trained model-\n",
    "                outputs = model(images)\n",
    "                _, y_pred = torch.max(outputs, 1)\n",
    "                \n",
    "                # Compute validation loss-\n",
    "                J_val = loss(outputs, labels)\n",
    "                \n",
    "                running_loss_val += J_val.item() * labels.size(0)\n",
    "    \n",
    "                # Total number of labels-\n",
    "                total += labels.size(0)\n",
    "\n",
    "                # Total number of correct predictions-\n",
    "                correct += (y_pred == labels).sum()\n",
    "                \n",
    "                tepoch.set_postfix(\n",
    "                    val_loss = running_loss_val / len(test_dataset),\n",
    "                    val_acc = 100 * (correct.cpu().numpy() / total)\n",
    "                )\n",
    "            \n",
    "        \n",
    "    # return (running_loss_val, correct, total)\n",
    "    val_loss = running_loss_val / len(test_dataset)\n",
    "    val_acc = (correct / total) * 100\n",
    "\n",
    "    return val_loss, val_acc.cpu().numpy()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f14b2e2b-0087-4702-91de-814cbe26a8e4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "094df51a-2763-49c5-8b4b-9510ee2489fd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Python3 dict to contain training metrics-\n",
    "training_history_lr_scheduler = {}\n",
    "\n",
    "# Initialize parameters saving 'best' models-\n",
    "best_val_acc = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "95f955f0-ca1c-422a-815b-a3189fdd4874",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "8c5d8aaf-554d-4910-a6c4-5752b154beaa",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.07batch/s, accuracy=15.6, loss=2.25]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.40batch/s, val_acc=21.7, val_loss=2.33]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 1 training loss = 2.2529, training accuracy = 15.61%, val_loss = 2.3296, val_accuracy = 21.66% LR = 0.0077 & beta = 5.323336\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:33<00:00,  2.10batch/s, accuracy=25.1, loss=2.13]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=36.5, val_loss=1.77]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 2 training loss = 2.1254, training accuracy = 25.14%, val_loss = 1.7741, val_accuracy = 36.55% LR = 0.0154 & beta = 5.157431\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=33.5, loss=1.88]\n",
      "Validation: : 100%|███████████████████████████████████| 40/40 [00:04<00:00,  9.36batch/s, val_acc=39, val_loss=1.71]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 3 training loss = 1.8773, training accuracy = 33.51%, val_loss = 1.7106, val_accuracy = 38.95% LR = 0.0231 & beta = 4.915476\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=39.3, loss=1.69]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.31batch/s, val_acc=46.6, val_loss=1.45]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 4 training loss = 1.6920, training accuracy = 39.27%, val_loss = 1.4487, val_accuracy = 46.57% LR = 0.0308 & beta = 4.629663\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=45.2, loss=1.54]\n",
      "Validation: : 100%|████████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=51, val_loss=1.4]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 5 training loss = 1.5357, training accuracy = 45.20%, val_loss = 1.3976, val_accuracy = 50.96% LR = 0.0386 & beta = 4.303523\n",
      "\n",
      "Saving model with highest val acc = 50.9600%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|█████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=50.1, loss=1.4]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=58.2, val_loss=1.21]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 6 training loss = 1.3976, training accuracy = 50.11%, val_loss = 1.2062, val_accuracy = 58.25% LR = 0.0463 & beta = 3.902002\n",
      "\n",
      "Saving model with highest val acc = 58.2500%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=54.6, loss=1.29]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=53.5, val_loss=1.68]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 7 training loss = 1.2855, training accuracy = 54.56%, val_loss = 1.6827, val_accuracy = 53.54% LR = 0.0540 & beta = 3.400167\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=59.5, loss=1.15]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=66.3, val_loss=0.961]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 8 training loss = 1.1493, training accuracy = 59.53%, val_loss = 0.9612, val_accuracy = 66.35% LR = 0.0617 & beta = 2.954192\n",
      "\n",
      "Saving model with highest val acc = 66.3500%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=62.7, loss=1.06]\n",
      "Validation: : 100%|██████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=68.9, val_loss=0.9]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 9 training loss = 1.0565, training accuracy = 62.69%, val_loss = 0.9004, val_accuracy = 68.88% LR = 0.0694 & beta = 2.677884\n",
      "\n",
      "Saving model with highest val acc = 68.8800%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=65.3, loss=0.995]\n",
      "Validation: : 100%|██████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=69, val_loss=0.887]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 10 training loss = 0.9949, training accuracy = 65.26%, val_loss = 0.8874, val_accuracy = 69.04% LR = 0.0772 & beta = 2.682029\n",
      "\n",
      "Saving model with highest val acc = 69.0400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=67.6, loss=0.933]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.31batch/s, val_acc=69.5, val_loss=0.868]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 11 training loss = 0.9328, training accuracy = 67.56%, val_loss = 0.8676, val_accuracy = 69.54% LR = 0.0849 & beta = 2.784408\n",
      "\n",
      "Saving model with highest val acc = 69.5400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=69.7, loss=0.872]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=75.5, val_loss=0.716]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 12 training loss = 0.8723, training accuracy = 69.73%, val_loss = 0.7157, val_accuracy = 75.53% LR = 0.0926 & beta = 2.909071\n",
      "\n",
      "Saving model with highest val acc = 75.5300%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=70.6, loss=0.847]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=72.2, val_loss=0.818]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 13 training loss = 0.8471, training accuracy = 70.57%, val_loss = 0.8180, val_accuracy = 72.24% LR = 0.1000 & beta = 3.193128\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=71.8, loss=0.814]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=73.1, val_loss=0.794]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 14 training loss = 0.8137, training accuracy = 71.77%, val_loss = 0.7939, val_accuracy = 73.09% LR = 0.1000 & beta = 3.481975\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=72.9, loss=0.776]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=76.8, val_loss=0.67]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 15 training loss = 0.7756, training accuracy = 72.88%, val_loss = 0.6702, val_accuracy = 76.83% LR = 0.1000 & beta = 3.706782\n",
      "\n",
      "Saving model with highest val acc = 76.8300%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=74.2, loss=0.745]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=77.2, val_loss=0.658]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 16 training loss = 0.7454, training accuracy = 74.21%, val_loss = 0.6584, val_accuracy = 77.25% LR = 0.1000 & beta = 4.037765\n",
      "\n",
      "Saving model with highest val acc = 77.2500%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=75.2, loss=0.714]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=79.7, val_loss=0.592]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 17 training loss = 0.7143, training accuracy = 75.21%, val_loss = 0.5918, val_accuracy = 79.70% LR = 0.1000 & beta = 4.359786\n",
      "\n",
      "Saving model with highest val acc = 79.7000%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=75.3, loss=0.708]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=74.5, val_loss=0.752]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 18 training loss = 0.7084, training accuracy = 75.28%, val_loss = 0.7517, val_accuracy = 74.45% LR = 0.1000 & beta = 4.644651\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=75.9, loss=0.689]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.31batch/s, val_acc=79.9, val_loss=0.586]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 19 training loss = 0.6892, training accuracy = 75.88%, val_loss = 0.5864, val_accuracy = 79.91% LR = 0.1000 & beta = 4.978436\n",
      "\n",
      "Saving model with highest val acc = 79.9100%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=76.7, loss=0.667]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=83.9, val_loss=0.475]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 20 training loss = 0.6673, training accuracy = 76.67%, val_loss = 0.4753, val_accuracy = 83.91% LR = 0.0100 & beta = 5.235280\n",
      "\n",
      "Saving model with highest val acc = 83.9100%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=82.8, loss=0.488]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=87.4, val_loss=0.383]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 21 training loss = 0.4879, training accuracy = 82.79%, val_loss = 0.3834, val_accuracy = 87.35% LR = 0.0100 & beta = 5.254888\n",
      "\n",
      "Saving model with highest val acc = 87.3500%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=84.9, loss=0.433]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=87.5, val_loss=0.364]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 22 training loss = 0.4329, training accuracy = 84.89%, val_loss = 0.3635, val_accuracy = 87.49% LR = 0.0100 & beta = 5.259077\n",
      "\n",
      "Saving model with highest val acc = 87.4900%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=85.8, loss=0.411]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=88.2, val_loss=0.354]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 23 training loss = 0.4108, training accuracy = 85.81%, val_loss = 0.3540, val_accuracy = 88.16% LR = 0.0100 & beta = 5.260085\n",
      "\n",
      "Saving model with highest val acc = 88.1600%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=86.3, loss=0.394]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=88.9, val_loss=0.33]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 24 training loss = 0.3936, training accuracy = 86.32%, val_loss = 0.3302, val_accuracy = 88.86% LR = 0.0100 & beta = 5.265076\n",
      "\n",
      "Saving model with highest val acc = 88.8600%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=86.8, loss=0.381]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=88.9, val_loss=0.329]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 25 training loss = 0.3812, training accuracy = 86.84%, val_loss = 0.3286, val_accuracy = 88.94% LR = 0.0100 & beta = 5.276586\n",
      "\n",
      "Saving model with highest val acc = 88.9400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=87.1, loss=0.368]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=89.1, val_loss=0.323]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 26 training loss = 0.3681, training accuracy = 87.14%, val_loss = 0.3231, val_accuracy = 89.13% LR = 0.0100 & beta = 5.284295\n",
      "\n",
      "Saving model with highest val acc = 89.1300%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=87.6, loss=0.357]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=89.1, val_loss=0.317]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 27 training loss = 0.3573, training accuracy = 87.56%, val_loss = 0.3171, val_accuracy = 89.08% LR = 0.0100 & beta = 5.297175\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|█████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=88, loss=0.341]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.26batch/s, val_acc=89.4, val_loss=0.31]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 28 training loss = 0.3413, training accuracy = 88.02%, val_loss = 0.3103, val_accuracy = 89.44% LR = 0.0100 & beta = 5.317672\n",
      "\n",
      "Saving model with highest val acc = 89.4400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=88.6, loss=0.328]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=89.3, val_loss=0.311]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 29 training loss = 0.3280, training accuracy = 88.64%, val_loss = 0.3111, val_accuracy = 89.28% LR = 0.0100 & beta = 5.335474\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=88.8, loss=0.324]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=89.5, val_loss=0.303]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 30 training loss = 0.3236, training accuracy = 88.84%, val_loss = 0.3027, val_accuracy = 89.54% LR = 0.0100 & beta = 5.349105\n",
      "\n",
      "Saving model with highest val acc = 89.5400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=89.2, loss=0.312]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=89.4, val_loss=0.312]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 31 training loss = 0.3122, training accuracy = 89.15%, val_loss = 0.3125, val_accuracy = 89.43% LR = 0.0100 & beta = 5.367851\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=89.3, loss=0.31]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=89.6, val_loss=0.31]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 32 training loss = 0.3097, training accuracy = 89.33%, val_loss = 0.3098, val_accuracy = 89.60% LR = 0.0100 & beta = 5.393928\n",
      "\n",
      "Saving model with highest val acc = 89.6000%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=89.7, loss=0.299]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=89.2, val_loss=0.309]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 33 training loss = 0.2990, training accuracy = 89.71%, val_loss = 0.3086, val_accuracy = 89.18% LR = 0.0100 & beta = 5.424795\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=89.8, loss=0.295]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=89.8, val_loss=0.306]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 34 training loss = 0.2951, training accuracy = 89.81%, val_loss = 0.3064, val_accuracy = 89.85% LR = 0.0100 & beta = 5.449564\n",
      "\n",
      "Saving model with highest val acc = 89.8500%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=90.2, loss=0.284]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=89.8, val_loss=0.306]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 35 training loss = 0.2842, training accuracy = 90.15%, val_loss = 0.3061, val_accuracy = 89.79% LR = 0.0100 & beta = 5.479353\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=90.4, loss=0.277]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=89.8, val_loss=0.304]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 36 training loss = 0.2770, training accuracy = 90.41%, val_loss = 0.3043, val_accuracy = 89.75% LR = 0.0100 & beta = 5.525251\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=90.6, loss=0.271]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=89.8, val_loss=0.302]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 37 training loss = 0.2706, training accuracy = 90.64%, val_loss = 0.3017, val_accuracy = 89.75% LR = 0.0100 & beta = 5.555652\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.1, loss=0.263]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=89.3, val_loss=0.323]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 38 training loss = 0.2628, training accuracy = 91.07%, val_loss = 0.3226, val_accuracy = 89.30% LR = 0.0100 & beta = 5.593637\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.1, loss=0.263]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=89.8, val_loss=0.311]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 39 training loss = 0.2626, training accuracy = 91.08%, val_loss = 0.3108, val_accuracy = 89.76% LR = 0.0100 & beta = 5.629722\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.2, loss=0.254]\n",
      "Validation: : 100%|██████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=90, val_loss=0.298]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 40 training loss = 0.2540, training accuracy = 91.22%, val_loss = 0.2977, val_accuracy = 90.03% LR = 0.0100 & beta = 5.666465\n",
      "\n",
      "Saving model with highest val acc = 90.0300%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.6, loss=0.246]\n",
      "Validation: : 100%|██████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=90, val_loss=0.307]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 41 training loss = 0.2456, training accuracy = 91.61%, val_loss = 0.3072, val_accuracy = 89.99% LR = 0.0100 & beta = 5.702801\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.7, loss=0.246]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=89.6, val_loss=0.305]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 42 training loss = 0.2458, training accuracy = 91.70%, val_loss = 0.3046, val_accuracy = 89.58% LR = 0.0100 & beta = 5.741313\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.9, loss=0.238]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=89.9, val_loss=0.303]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 43 training loss = 0.2384, training accuracy = 91.86%, val_loss = 0.3031, val_accuracy = 89.94% LR = 0.0100 & beta = 5.779944\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|█████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=92, loss=0.236]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.31batch/s, val_acc=89.1, val_loss=0.324]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 44 training loss = 0.2359, training accuracy = 91.96%, val_loss = 0.3237, val_accuracy = 89.13% LR = 0.0100 & beta = 5.823212\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=91.9, loss=0.236]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=90.3, val_loss=0.292]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 45 training loss = 0.2357, training accuracy = 91.87%, val_loss = 0.2918, val_accuracy = 90.35% LR = 0.0010 & beta = 5.853659\n",
      "\n",
      "Saving model with highest val acc = 90.3500%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=93.5, loss=0.192]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.30batch/s, val_acc=90.8, val_loss=0.275]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 46 training loss = 0.1925, training accuracy = 93.53%, val_loss = 0.2750, val_accuracy = 90.83% LR = 0.0010 & beta = 5.856497\n",
      "\n",
      "Saving model with highest val acc = 90.8300%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=94.2, loss=0.177]\n",
      "Validation: : 100%|██████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=91, val_loss=0.272]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 47 training loss = 0.1770, training accuracy = 94.22%, val_loss = 0.2716, val_accuracy = 91.04% LR = 0.0010 & beta = 5.857751\n",
      "\n",
      "Saving model with highest val acc = 91.0400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=94.5, loss=0.172]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.3, val_loss=0.264]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 48 training loss = 0.1718, training accuracy = 94.50%, val_loss = 0.2643, val_accuracy = 91.32% LR = 0.0010 & beta = 5.859433\n",
      "\n",
      "Saving model with highest val acc = 91.3200%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=94.6, loss=0.166]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.2, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 49 training loss = 0.1661, training accuracy = 94.56%, val_loss = 0.2669, val_accuracy = 91.17% LR = 0.0010 & beta = 5.860623\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=94.7, loss=0.162]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.3, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 50 training loss = 0.1622, training accuracy = 94.72%, val_loss = 0.2665, val_accuracy = 91.26% LR = 0.0010 & beta = 5.861127\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=94.8, loss=0.158]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.4, val_loss=0.266]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 51 training loss = 0.1582, training accuracy = 94.80%, val_loss = 0.2658, val_accuracy = 91.42% LR = 0.0010 & beta = 5.862113\n",
      "\n",
      "Saving model with highest val acc = 91.4200%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=94.9, loss=0.156]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.4, val_loss=0.265]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 52 training loss = 0.1561, training accuracy = 94.93%, val_loss = 0.2654, val_accuracy = 91.44% LR = 0.0010 & beta = 5.863811\n",
      "\n",
      "Saving model with highest val acc = 91.4400%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=94.9, loss=0.153]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.3, val_loss=0.264]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 53 training loss = 0.1530, training accuracy = 94.92%, val_loss = 0.2645, val_accuracy = 91.32% LR = 0.0010 & beta = 5.863107\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=95.2, loss=0.148]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.4, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 54 training loss = 0.1481, training accuracy = 95.16%, val_loss = 0.2691, val_accuracy = 91.38% LR = 0.0010 & beta = 5.864272\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.2, loss=0.15]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.4, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 55 training loss = 0.1497, training accuracy = 95.18%, val_loss = 0.2689, val_accuracy = 91.43% LR = 0.0010 & beta = 5.865068\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.3, loss=0.146]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.3, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 56 training loss = 0.1464, training accuracy = 95.33%, val_loss = 0.2669, val_accuracy = 91.34% LR = 0.0010 & beta = 5.865669\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.3, loss=0.146]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.3, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 57 training loss = 0.1459, training accuracy = 95.29%, val_loss = 0.2682, val_accuracy = 91.27% LR = 0.0010 & beta = 5.866251\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=95.5, loss=0.139]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.3, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 58 training loss = 0.1392, training accuracy = 95.49%, val_loss = 0.2692, val_accuracy = 91.30% LR = 0.0010 & beta = 5.866070\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=95.5, loss=0.144]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.4, val_loss=0.265]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 59 training loss = 0.1436, training accuracy = 95.46%, val_loss = 0.2654, val_accuracy = 91.40% LR = 0.0010 & beta = 5.866460\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|████████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=95.6, loss=0.14]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.3, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 60 training loss = 0.1401, training accuracy = 95.56%, val_loss = 0.2691, val_accuracy = 91.28% LR = 0.0010 & beta = 5.866505\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.6, loss=0.136]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.36batch/s, val_acc=91.2, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 61 training loss = 0.1359, training accuracy = 95.60%, val_loss = 0.2687, val_accuracy = 91.25% LR = 0.0010 & beta = 5.865776\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.7, loss=0.135]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.4, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 62 training loss = 0.1354, training accuracy = 95.71%, val_loss = 0.2692, val_accuracy = 91.42% LR = 0.0010 & beta = 5.865740\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=95.6, loss=0.134]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.4, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 63 training loss = 0.1336, training accuracy = 95.62%, val_loss = 0.2682, val_accuracy = 91.42% LR = 0.0010 & beta = 5.865839\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.7, loss=0.134]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=91.5, val_loss=0.27]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 64 training loss = 0.1337, training accuracy = 95.70%, val_loss = 0.2700, val_accuracy = 91.46% LR = 0.0010 & beta = 5.865566\n",
      "\n",
      "Saving model with highest val acc = 91.4600%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|█████████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=96, loss=0.127]\n",
      "Validation: : 100%|█████████████████████████████████| 40/40 [00:04<00:00,  9.44batch/s, val_acc=91.2, val_loss=0.27]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 65 training loss = 0.1266, training accuracy = 95.98%, val_loss = 0.2703, val_accuracy = 91.18% LR = 0.0010 & beta = 5.866371\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.9, loss=0.131]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.3, val_loss=0.271]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 66 training loss = 0.1309, training accuracy = 95.88%, val_loss = 0.2709, val_accuracy = 91.32% LR = 0.0010 & beta = 5.866257\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.9, loss=0.127]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.2, val_loss=0.271]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 67 training loss = 0.1268, training accuracy = 95.88%, val_loss = 0.2712, val_accuracy = 91.25% LR = 0.0010 & beta = 5.866765\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=95.9, loss=0.128]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.4, val_loss=0.269]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 68 training loss = 0.1280, training accuracy = 95.95%, val_loss = 0.2693, val_accuracy = 91.41% LR = 0.0010 & beta = 5.867671\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.1, loss=0.123]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.2, val_loss=0.271]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 69 training loss = 0.1228, training accuracy = 96.09%, val_loss = 0.2708, val_accuracy = 91.19% LR = 0.0010 & beta = 5.867098\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=96.1, loss=0.123]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.5, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 70 training loss = 0.1232, training accuracy = 96.12%, val_loss = 0.2682, val_accuracy = 91.45% LR = 0.0001 & beta = 5.867257\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.3, loss=0.118]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.6, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 71 training loss = 0.1175, training accuracy = 96.29%, val_loss = 0.2681, val_accuracy = 91.56% LR = 0.0001 & beta = 5.867233\n",
      "\n",
      "Saving model with highest val acc = 91.5600%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.2, loss=0.122]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.5, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 72 training loss = 0.1216, training accuracy = 96.24%, val_loss = 0.2674, val_accuracy = 91.50% LR = 0.0001 & beta = 5.867229\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.2, loss=0.122]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.36batch/s, val_acc=91.5, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 73 training loss = 0.1219, training accuracy = 96.15%, val_loss = 0.2683, val_accuracy = 91.54% LR = 0.0001 & beta = 5.867147\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.4, loss=0.117]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=91.5, val_loss=0.266]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 74 training loss = 0.1172, training accuracy = 96.40%, val_loss = 0.2658, val_accuracy = 91.53% LR = 0.0001 & beta = 5.867158\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.2, loss=0.119]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.5, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 75 training loss = 0.1187, training accuracy = 96.21%, val_loss = 0.2684, val_accuracy = 91.52% LR = 0.0001 & beta = 5.867083\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=96.3, loss=0.117]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.5, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 76 training loss = 0.1166, training accuracy = 96.34%, val_loss = 0.2677, val_accuracy = 91.45% LR = 0.0001 & beta = 5.867015\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.3, loss=0.117]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.5, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 77 training loss = 0.1170, training accuracy = 96.28%, val_loss = 0.2683, val_accuracy = 91.50% LR = 0.0001 & beta = 5.866958\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=96.4, loss=0.114]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.37batch/s, val_acc=91.6, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 78 training loss = 0.1138, training accuracy = 96.39%, val_loss = 0.2683, val_accuracy = 91.59% LR = 0.0001 & beta = 5.866834\n",
      "\n",
      "Saving model with highest val acc = 91.5900%\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.3, loss=0.118]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.5, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 79 training loss = 0.1175, training accuracy = 96.32%, val_loss = 0.2682, val_accuracy = 91.49% LR = 0.0001 & beta = 5.866755\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.4, loss=0.114]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.6, val_loss=0.268]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 80 training loss = 0.1138, training accuracy = 96.45%, val_loss = 0.2677, val_accuracy = 91.58% LR = 0.0001 & beta = 5.866787\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.4, loss=0.114]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.5, val_loss=0.266]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 81 training loss = 0.1142, training accuracy = 96.43%, val_loss = 0.2664, val_accuracy = 91.53% LR = 0.0001 & beta = 5.866889\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.3, loss=0.119]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.36batch/s, val_acc=91.5, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 82 training loss = 0.1193, training accuracy = 96.28%, val_loss = 0.2672, val_accuracy = 91.47% LR = 0.0001 & beta = 5.866879\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:34<00:00,  2.08batch/s, accuracy=96.4, loss=0.115]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.33batch/s, val_acc=91.6, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 83 training loss = 0.1154, training accuracy = 96.41%, val_loss = 0.2668, val_accuracy = 91.56% LR = 0.0001 & beta = 5.866838\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.5, loss=0.112]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.35batch/s, val_acc=91.6, val_loss=0.267]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 84 training loss = 0.1117, training accuracy = 96.54%, val_loss = 0.2670, val_accuracy = 91.56% LR = 0.0001 & beta = 5.866815\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: : 100%|███████████████████████████████████| 196/196 [01:33<00:00,  2.09batch/s, accuracy=96.4, loss=0.115]\n",
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.34batch/s, val_acc=91.6, val_loss=0.268]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "epoch: 85 training loss = 0.1153, training accuracy = 96.42%, val_loss = 0.2682, val_accuracy = 91.58% LR = 0.0001 & beta = 5.866677\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for epoch in range(num_epochs):\n",
    "\n",
    "    # Train and validate model for 1 epoch-\n",
    "    train_loss, train_acc = train_model_progress(\n",
    "        model = model, train_loader = train_loader,\n",
    "        train_dataset = train_dataset\n",
    "    )\n",
    "    val_loss, val_acc = test_model_progress(\n",
    "        model = model, test_loader = val_loader,\n",
    "        test_dataset = val_dataset\n",
    "    )\n",
    "    \n",
    "    curr_lr = optimizer.param_groups[0]['lr']\n",
    "    curr_beta = model.swish_fn.beta.detach().item()\n",
    "    \n",
    "    print(f\"\\nepoch: {epoch + 1} training loss = {train_loss:.4f}, \"\n",
    "          f\"training accuracy = {train_acc:.2f}%, val_loss = {val_loss:.4f}\"\n",
    "          f\", val_accuracy = {val_acc:.2f}% \"\n",
    "          f\"LR = {curr_lr:.4f} & beta = {curr_beta:.6f}\\n\")\n",
    "    \n",
    "    training_history_lr_scheduler[epoch + 1] = {\n",
    "        'loss': train_loss, 'acc': train_acc,\n",
    "        'val_loss': val_loss, 'val_acc': val_acc,\n",
    "        'lr': curr_lr, 'beta': curr_beta\n",
    "    }\n",
    "    \n",
    "    # Save best weights achieved until now-\n",
    "    if (val_acc > best_val_acc):\n",
    "        # update 'best_val_loss' variable to lowest loss encountered so far-\n",
    "        best_val_acc = val_acc\n",
    "\n",
    "        print(f\"Saving model with highest val acc = {val_acc:.4f}%\\n\")\n",
    "        \n",
    "        # Save trained model with 'best' validation accuracy-\n",
    "        torch.save(model.state_dict(), \"ResNet50_best_valacc_model.pth\")\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7d9d23ca-5658-4cc7-aef6-00f6ee06b8f3",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "f7cf3533-f2de-4659-9e8a-bb1c59c37bb3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save training metrics as Python3 history for later analysis-\n",
    "with open(\"ResNet50_swish_training_history.pkl\", \"wb\") as file:\n",
    "    pickle.dump(training_history_lr_scheduler, file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6581b739-ffff-49fc-94a0-1b76e05361a8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "cef6b168-5e0a-41ee-8697-47d87ef24a6c",
   "metadata": {},
   "outputs": [],
   "source": [
    "del model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e6d2e412-4d88-461e-8a4d-ae595d2a46f4",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5a3aac2c-2346-4c0d-9403-953ace452750",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "655f8ad7-9c98-4eb8-9004-a101f24a3a8d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (9, 7))\n",
    "plt.plot(list(training_history_lr_scheduler.keys()), [training_history_lr_scheduler[k]['acc'] for k in training_history_lr_scheduler.keys()], label = 'training acc')\n",
    "plt.plot(list(training_history_lr_scheduler.keys()), [training_history_lr_scheduler[k]['val_acc'] for k in training_history_lr_scheduler.keys()], label = 'val acc')\n",
    "plt.title(\"ResNet-50: Training Accuracy\")\n",
    "plt.xlabel(\"epochs\")\n",
    "plt.ylabel(\"accuracy (%)\")\n",
    "plt.legend(loc = 'best')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "95ee8bc7-b387-48f4-a29c-089d8f21a0d0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (9, 7))\n",
    "plt.plot(list(training_history_lr_scheduler.keys()), [training_history_lr_scheduler[k]['loss'] for k in training_history_lr_scheduler.keys()], label = 'training loss')\n",
    "plt.plot(list(training_history_lr_scheduler.keys()), [training_history_lr_scheduler[k]['val_loss'] for k in training_history_lr_scheduler.keys()], label = 'val loss')\n",
    "plt.xlabel(\"epochs\")\n",
    "plt.ylabel(\"loss\")\n",
    "plt.legend(loc = 'best')\n",
    "plt.title(\"ResNet-50: Training Loss\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "2aedb3e1-d3c2-490f-90d8-159aa5ada459",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (9, 7))\n",
    "plt.plot(list(training_history_lr_scheduler.keys()), [training_history_lr_scheduler[k]['lr'] for k in training_history_lr_scheduler.keys()])\n",
    "plt.xlabel(\"epochs\")\n",
    "plt.ylabel(\"lr\")\n",
    "plt.title(\"ResNet-50: Learning-Rate\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "3d704cac-4aa3-4b0f-bb43-7d6e186e810b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize = (9, 7))\n",
    "plt.plot(list(training_history_lr_scheduler.keys()), [training_history_lr_scheduler[k]['beta'] for k in training_history_lr_scheduler.keys()])\n",
    "plt.xlabel(\"epochs\")\n",
    "plt.ylabel(\"beta\")\n",
    "plt.title(\"ResNet-50: Swish (beta) activation function\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5d1f4698-e0db-456d-9aec-c75934d76181",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "258cd1d1-2692-427a-9df9-d164c1f04f38",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "highest val acc = 91.59%  is obtained for beta = 5.8668\n"
     ]
    }
   ],
   "source": [
    "# max([train_history[k]['val_acc'] for k in train_history.keys()])\n",
    "max_val_acc = 0\n",
    "max_epoch = 0\n",
    "\n",
    "for e in training_history_lr_scheduler.keys():\n",
    "    if training_history_lr_scheduler[e]['val_acc'] > max_val_acc:\n",
    "        max_val_acc = training_history_lr_scheduler[e]['val_acc']\n",
    "        max_epoch = e\n",
    "\n",
    "print(f\"highest val acc = {training_history_lr_scheduler[max_epoch]['val_acc']:.2f}%\",\n",
    "      f\" is obtained for beta = {training_history_lr_scheduler[max_epoch]['beta']:.4f}\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f3a42f25-fbd2-4efc-a7a8-befc11e69d63",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "edf04050-fcf5-4a64-b89d-fbd643fda9e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize and load 'best' trained parameters from above-\n",
    "best_model = ResNet50()\n",
    "best_model.load_state_dict(torch.load(\"ResNet50_best_valacc_model.pth\"))\n",
    "\n",
    "# Move to GPU device-\n",
    "best_model.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "dec81d42-73d2-4e83-94c1-f9e6df0674e0",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Validation: : 100%|████████████████████████████████| 40/40 [00:04<00:00,  9.32batch/s, val_acc=91.6, val_loss=0.268]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ResNet-50 'best' trained validation metrics:  loss = 0.2683 & acc = 91.59%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# Get validation metrics of 'best' trained model-\n",
    "val_loss, val_acc = test_model_progress(\n",
    "    model = best_model, test_loader = val_loader,\n",
    "    test_dataset = val_dataset\n",
    ")\n",
    "\n",
    "print(\"ResNet-50 'best' trained validation metrics: \",\n",
    "      f\"loss = {val_loss:.4f} & acc = {val_acc:.2f}%\"\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3810040b-e519-40aa-8bab-b427ecdcaf5f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fcb37708-683b-4bf3-8357-ca9d870f0d97",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}